Progressive platelet activation with storage: evidence for shortened survival of activated platelets after transfusion H.M. RINDER,M. MURPHY,J.G. MITCHELL,J. STOCKS,K.A. AULT,AND R.S. HILLMAN Platelets are known to become activated durin storage, but it is unclear whether such activation affects recovery or survival after Jatelet concentrate (PC) transfusion With the use of flow cytometry to determine the percentage of platelets expressing the alphagranule membrane protein 140 (GMP-140), a known adhesive ligand a p aring on the platelet surface after activation, several studies were conducted. These p“ nvestigations evaluated 1) the occurrence of signiflcant platelet activation over time in PCs (n = 46) stored under standard blood bank conditions; 2) the correlation between platelet activation and platelet recovey In normal subjects after PC storage (n = 12), as assessed by the recovery of Indium-labeled platelets; and 3) the recovery of activated and unactivated platelets in thrombocytopenic cancer patients transfused with standard PCs (n = 11). It was determined 1) that an increasing duration of storage of PC was associated with Increasing platelet activation as measured by the percenta e of platelets expressing GMP-140, progressin from a mean of 4 2 2 percent (SOY on the day of collection to a mean of 25 k #percent by 5 days of storage: 2) that, in normal subjects, posttransfusion recovery of autologous platelets stored for 2 to 4 days and then labeled with Iniii was Inversely correlated with the percentage of activated platelets in the transfused PC (r = -0.55, p = 0.05); and 3) that, when thrombocytopenic patients were transfused with standard PCs, the recovery of the activated platelets In the transfused PCs averaged only 38 2 15 percent of the number predicted by the absolute platelet increment. It c a n be concluded that significant platelet activation occurs with standard platelet stora e over 5 days and that activated platelets that express GMP-140 are preferentiafy cleared from the circulation after transfusion. TRANSFUSION 1991;31:409-414. Abbrev!atlonr: FITC = fluorerceln irothlocyanate;QMP-140 = alphagranule membrane proteln-140; PC(8) = platelet concentrate(a); PE = phycoetythrin; PPP = platelet-poor plasma.
THETRANSFUSIONOF platelet concentrates (PO) is an essential supportive therapy in patients with severe thrombocytopenia, especially those undergoing intensive chemotherapy. Transfusion efficacy is determined from the increment in the platelet count after transfusion and the impact on clinical bleeding. Therefore, relatively normal function and adequate recovery of stored platelets following transfusion are essential. Storage temperature, pH, agitation, and contamination with white cells are variables that are known to affect stored platelet function and recovery.’” Even under optimal conditions, platelet function has been demonstrated to decline over time with storage,c6 and, with increasing storage time as an independent variable, increments in platelet count after transfusion fall.2*7*8 We and othersg observed that platelets in storage undergo alpha-granule release, as determined by in-
creased expression of alpha-granule membrane protein 140 (GMP-140) on the platelet surface. GMP-140 appears to function as an adhesive platelet ligand; thus, it is possible that platelets expressing this molecule would be preferentially cleared from the circulation. We used a monoclonal antibody specific for GMP-140, which is expressed only on activated platelets,’”12 and a flow cytometric assay to determine directly the percentage of activated (GMP-140-expressing) platelets in stored PCs and in patient whole blood. This study was designed to determine 1) whether increasing storage time is associated with increases in the percentage of platelets expressing GMP-140; 2) whether platelet recovery in normal subjects after transfusion of indium (Inl’’)-labeled autologous PCs correlates with the percentage of platelets expressing GMP-140 in the transfused autologous PCs; and 3) whether stored platelets expressing GMP-140 can be recovered in the circulation of thrombocytopenic patients after transfusion or whether they might have been preferentially cleared. Our studies demonstrate a progressive activation of stored platelets with time and suggest a direct correlation between platelet activation, as determined by GMP-140
From the Departments of Medicine and Pathology, Maine Medical Center, and the Maine Cytometry Research Institute. Portland, Maine. Supported in part by a grant from the Maine Chapter of the American Heart Association. Received for publiation June 27.1989; revision received November 9,1990,and accepted k r n b e r 9,1990.
409
410
TRANSFUSION Vd. 31. No. 5-1991
RINDER ET AL.
expression, and a failure of platelets to circulate at 1 hour after transfusion.
Materials and Methods All studies of human subjects were approved by the institutional review board of the Maine Medical Center.
Activation of platelets in blood bank-stomd PCs At the Maine Medical Center, we sterilely drew 50-pL samples from 46 individual randomdonor platelet units after 3 to 6 days of storage. Longitudinal studies of the same platelet unit were not carried out to avoid undue interference in the use of these platelets for patient transfusions. For this part of the study, we performed no further experiments on these platelet units or the patients who received them. We drew 15 samples of whole blood from blood donors for time zero (collection time) analysis. The above samples were fixed and analyzed for platelet activation as descn'bcd below.
Recovery of activated autologous platelets in noml subjects After obtaining informed consent, we performed recovery studies of autologous Inlll-labeled platelets in 12 normal volunteers using Stratton's technique,1s1s with the modification of an added CPD- or ACD-saline wash. Briefly, we collected donor blood in ACD or CPD using licensed plastic storage bags containing two integrally attached satellite bags (PG732, Fenwal Laboratories, Morton Grove, IL). Whole blood was centrifuged at 350 x g for 15 minutes, and we transferred the upper three-quarters of platelet-rich plasma to the platelet satellite bag and centrifuged it at 1300 X g for 15 minutes. All but 50 mL of the platelet-poor plasma (PPP) was extruded into the second satellite bag. We kept the platelet button stationaty for 1 hour at 22°C and then resuspended it in the remaining PPP by placing the platelet satellite bag on a linear platelet reciprocator with gentle horizontal agitation at 70 cpm at 22°C. No differences between CPD or ACD units were observed. Stored units were then sampled daily for the percentage of GMP-140 expression. Other variables, such as pH, were not determined. After 2 to 4 days of storage, we centrifuged the platelets at 1300 x g for 15 minutes, resuspended them in ACD or CPD, with saline, and incubated them with Inl"-oxine for 20 minutes. We then restored the PPP, centrifuged the platelets at 1300 x g for 15 minutes, and carefully decanted the supernatant to remove unbound indium and to estimate labeling efficiency. The final platelet pellet was resuspended in 8 to 10 mL of autologous plasma; aliquots were removed for counting standards and to measure GMP-140 expression, and the remainder was immediately injected into the autologous donor. The In111 labeling procedure itself resulted in a less than 5 to 10 percent increase in the proportion of GMP140-positiveplatelets. We then drew venous samples of whole blood without a tourniquet at 15,30, and 60 minutes for platelet count and In111counting. There was no significant difference in the indium values obtained at these time points, which indicated rapid equilibration of transfused platelets.16 We calculated the percentage of recovery at l hour by Ezekowitz's method.lS We did not perform survival studies.
Recovery of activated platelets in thrombaytopenic patients Eleven patients receiving therapeutic platelet transfusions gave informed consent. All patients were thrombocytopenic
secondary to chemotherapy. No patient was bleeding, febrile, i n f d at the time of the study, or receiving amphotericin B, and none had undergone splenectomy or had splenomegaly. Each patient was studied once. Random PCs (8 individual units of the same ABO and Rh type, which all outdated on the same day) were pooled for each platelet transfusion: PCs had been stored from 4 to 5 days. We collcctcd a 50-pL sample from the pooled concentrate immediately before transfusion to determine the percentage of activated platelets in the PC. We drew from the patient, without the use of a tourniquet, a 1-mL sample of venous blood for platelet count and GMP-140 measurement immediately before and then 60 minutes after the completion of the transfusion. We performed platelet counts on an automated counter (TW,Coulter Electronics, Hialeah, FL) and confiied them by manual counting. If activated (GMP-140-positive) and unactivated (GMP-140negative) platelets survive equally well after transfusion, then the percentage of activated platelets in a whole blood sample following transfusion may be calculated by the following equation: Predicted activated platelets (%) =
A x D
+B X
(C-D)
C
where A is the percentage of activated platelets in the pretransfusion whole blood sample; B is the percentage of activated platelets in the PC; C is the platelet count at 1 hour after transfusion; and D is the platelet count prior to transfusion. To determine the relative in vivo suryiyal of GMP-140-positive and GMP-140-negative heterologous platelets in actual thrombocytopenic patients, we compared this predicted value to the actual percentage of activated platelets in the patient sample 60 minutes after transfusion. If activated (GMP-140positive) platelets swvivc less well than unactivated (GMP140-negative) platelets, this predicted percentage of activated platelets will be higher than the measured percentage of activated (GMP-140-positive) platelets. If, on the other hand, the GMP-140-positive platelets slwive better, the predicted value will be lower than the measured value. Note that this calculation is independent of the corrected count increment, as it is based on the actual percentages of activated platelets in the PCS and in the patient.
Monoclonal antibodies A murine monoclonal antihdy, S12, which is dinded against GMP-140 and has been demonstrated to bind only to activated platelets,1s12 was provided by Dr. R. McEver (University of Oklahoma, Oklahoma City, OK) and was biotinylated by Dr. A. Jackson (Becton Dickinson Immunocytometry Systems, Mountain View, CA). The monoclonal antibody anti-GPIb (AMAC, Westbrook, ME)was oonjugated to fluorescein isothiocyanate (FITC) and employed as a platelet marker in whole blood samples as described previo~sly.~'*~"
Flow qtometric analysis We immediately placed 50 p L of whole blood (patient samples) or of PC in 1mL of 2-percent paraformaldehyde in phosphate-buffered saline and incubated it for 1 hour at 4°C. We than washed samples with Tyrode's buffer and incubated them in the standard fashion for 20 minutes at 22°C with a saturating concentration of biotinylated S12 and FITC-anti-GPIb, as described prcviou~ly.~~ The sample was washed again, and 20 pL of phycoerythrin (PE)conjugated avidin (Becton Dickinson Immunccytometry Systems) was added for 20 minutes' incubation. PEanjugated avidin specifically binds biotiny-
'IRANSFUSION
1991-Vd. 31. No. 5
411
ACTIVATION OF STORED PLATELETS
lated S12 to produce PE fluorescence on the surface of cells expressing GMP-140. We again washed the samples in "yrode's buffer and analyzed them in a flow cytometer (FACScan, Becton Dickinson Immunocytometry Systems). The instrument was Set to measure forward light scatter (i.e., a measurement of particle size), and FITC and PE fluorescence. We calibrated the instrument for forward light scatter by using 2- and 10-km beads (Duke Scientific Corp., Palo Alto, CA) and for fluorescence by using calibration beads (Calibrite, Becton Dickinson Immunocytometry Systems). We identified platelets in whole blood samples by setting the cytometer to detect FITC fluorescence; that is, only GPIbpositive particles were measured as platelets. We also performed gating by using the light scatter profile of platelets to separate them from the larger red and white cells, as detailed previo~sly,~'.~~ and from any platelet fragments. The level of background PE fluorescence was determined on a daily basis using PEconjugated avidin alone, without biotinylated S12. We periodically labeled ADP-stimulated whole blood samples and analyzed them in a manner identical to that of the study samples to confirm that the S12 antibody was functional. Analysis of 10,OOOplatelets was performed, and the percentage of activated platelets was determined by the proportion of platelets that expressed PE fluorescence higher than background (Fig. 1). Although an 80-percent decline in GPIb fluorescence was noted with maximal platelet activation, the threshold for FITC detection was low enough to include all platelets. Dual platelet labeling with GPIb and GPIIbma showtd no platelets which had entirely lost GPIb as a consequence of storage. control studits with GPIIbAIIa and S12 labtling shaved similar percentages of activation as did those with GPIb and
Four Day Old
S12 labeling. We noted no GPIb negative platelet-sized cells, which confirmed that only platelets were analyzed.
Results GMP-140 q m s i o n in blood bank-stored PCs The relationship of GMP-140 expression-that is, the percentage of platelets that are activated-and the length of platelet storage is shown in Fig. 2. There is an obvious increase in platelet activation with storage; when compared to freshly drawn whole blood, platelets in storage were significantly activated (p
THETRANSFUSIONOF platelet concentrates (PO) is an essential supportive therapy in patients with severe thrombocytopenia, especially those undergoing intensive chemotherapy. Transfusion efficacy is determined from the increment in the platelet count after transfusion and the impact on clinical bleeding. Therefore, relatively normal function and adequate recovery of stored platelets following transfusion are essential. Storage temperature, pH, agitation, and contamination with white cells are variables that are known to affect stored platelet function and recovery.’” Even under optimal conditions, platelet function has been demonstrated to decline over time with storage,c6 and, with increasing storage time as an independent variable, increments in platelet count after transfusion fall.2*7*8 We and othersg observed that platelets in storage undergo alpha-granule release, as determined by in-
creased expression of alpha-granule membrane protein 140 (GMP-140) on the platelet surface. GMP-140 appears to function as an adhesive platelet ligand; thus, it is possible that platelets expressing this molecule would be preferentially cleared from the circulation. We used a monoclonal antibody specific for GMP-140, which is expressed only on activated platelets,’”12 and a flow cytometric assay to determine directly the percentage of activated (GMP-140-expressing) platelets in stored PCs and in patient whole blood. This study was designed to determine 1) whether increasing storage time is associated with increases in the percentage of platelets expressing GMP-140; 2) whether platelet recovery in normal subjects after transfusion of indium (Inl’’)-labeled autologous PCs correlates with the percentage of platelets expressing GMP-140 in the transfused autologous PCs; and 3) whether stored platelets expressing GMP-140 can be recovered in the circulation of thrombocytopenic patients after transfusion or whether they might have been preferentially cleared. Our studies demonstrate a progressive activation of stored platelets with time and suggest a direct correlation between platelet activation, as determined by GMP-140
From the Departments of Medicine and Pathology, Maine Medical Center, and the Maine Cytometry Research Institute. Portland, Maine. Supported in part by a grant from the Maine Chapter of the American Heart Association. Received for publiation June 27.1989; revision received November 9,1990,and accepted k r n b e r 9,1990.
409
410
TRANSFUSION Vd. 31. No. 5-1991
RINDER ET AL.
expression, and a failure of platelets to circulate at 1 hour after transfusion.
Materials and Methods All studies of human subjects were approved by the institutional review board of the Maine Medical Center.
Activation of platelets in blood bank-stomd PCs At the Maine Medical Center, we sterilely drew 50-pL samples from 46 individual randomdonor platelet units after 3 to 6 days of storage. Longitudinal studies of the same platelet unit were not carried out to avoid undue interference in the use of these platelets for patient transfusions. For this part of the study, we performed no further experiments on these platelet units or the patients who received them. We drew 15 samples of whole blood from blood donors for time zero (collection time) analysis. The above samples were fixed and analyzed for platelet activation as descn'bcd below.
Recovery of activated autologous platelets in noml subjects After obtaining informed consent, we performed recovery studies of autologous Inlll-labeled platelets in 12 normal volunteers using Stratton's technique,1s1s with the modification of an added CPD- or ACD-saline wash. Briefly, we collected donor blood in ACD or CPD using licensed plastic storage bags containing two integrally attached satellite bags (PG732, Fenwal Laboratories, Morton Grove, IL). Whole blood was centrifuged at 350 x g for 15 minutes, and we transferred the upper three-quarters of platelet-rich plasma to the platelet satellite bag and centrifuged it at 1300 X g for 15 minutes. All but 50 mL of the platelet-poor plasma (PPP) was extruded into the second satellite bag. We kept the platelet button stationaty for 1 hour at 22°C and then resuspended it in the remaining PPP by placing the platelet satellite bag on a linear platelet reciprocator with gentle horizontal agitation at 70 cpm at 22°C. No differences between CPD or ACD units were observed. Stored units were then sampled daily for the percentage of GMP-140 expression. Other variables, such as pH, were not determined. After 2 to 4 days of storage, we centrifuged the platelets at 1300 x g for 15 minutes, resuspended them in ACD or CPD, with saline, and incubated them with Inl"-oxine for 20 minutes. We then restored the PPP, centrifuged the platelets at 1300 x g for 15 minutes, and carefully decanted the supernatant to remove unbound indium and to estimate labeling efficiency. The final platelet pellet was resuspended in 8 to 10 mL of autologous plasma; aliquots were removed for counting standards and to measure GMP-140 expression, and the remainder was immediately injected into the autologous donor. The In111 labeling procedure itself resulted in a less than 5 to 10 percent increase in the proportion of GMP140-positiveplatelets. We then drew venous samples of whole blood without a tourniquet at 15,30, and 60 minutes for platelet count and In111counting. There was no significant difference in the indium values obtained at these time points, which indicated rapid equilibration of transfused platelets.16 We calculated the percentage of recovery at l hour by Ezekowitz's method.lS We did not perform survival studies.
Recovery of activated platelets in thrombaytopenic patients Eleven patients receiving therapeutic platelet transfusions gave informed consent. All patients were thrombocytopenic
secondary to chemotherapy. No patient was bleeding, febrile, i n f d at the time of the study, or receiving amphotericin B, and none had undergone splenectomy or had splenomegaly. Each patient was studied once. Random PCs (8 individual units of the same ABO and Rh type, which all outdated on the same day) were pooled for each platelet transfusion: PCs had been stored from 4 to 5 days. We collcctcd a 50-pL sample from the pooled concentrate immediately before transfusion to determine the percentage of activated platelets in the PC. We drew from the patient, without the use of a tourniquet, a 1-mL sample of venous blood for platelet count and GMP-140 measurement immediately before and then 60 minutes after the completion of the transfusion. We performed platelet counts on an automated counter (TW,Coulter Electronics, Hialeah, FL) and confiied them by manual counting. If activated (GMP-140-positive) and unactivated (GMP-140negative) platelets survive equally well after transfusion, then the percentage of activated platelets in a whole blood sample following transfusion may be calculated by the following equation: Predicted activated platelets (%) =
A x D
+B X
(C-D)
C
where A is the percentage of activated platelets in the pretransfusion whole blood sample; B is the percentage of activated platelets in the PC; C is the platelet count at 1 hour after transfusion; and D is the platelet count prior to transfusion. To determine the relative in vivo suryiyal of GMP-140-positive and GMP-140-negative heterologous platelets in actual thrombocytopenic patients, we compared this predicted value to the actual percentage of activated platelets in the patient sample 60 minutes after transfusion. If activated (GMP-140positive) platelets swvivc less well than unactivated (GMP140-negative) platelets, this predicted percentage of activated platelets will be higher than the measured percentage of activated (GMP-140-positive) platelets. If, on the other hand, the GMP-140-positive platelets slwive better, the predicted value will be lower than the measured value. Note that this calculation is independent of the corrected count increment, as it is based on the actual percentages of activated platelets in the PCS and in the patient.
Monoclonal antibodies A murine monoclonal antihdy, S12, which is dinded against GMP-140 and has been demonstrated to bind only to activated platelets,1s12 was provided by Dr. R. McEver (University of Oklahoma, Oklahoma City, OK) and was biotinylated by Dr. A. Jackson (Becton Dickinson Immunocytometry Systems, Mountain View, CA). The monoclonal antibody anti-GPIb (AMAC, Westbrook, ME)was oonjugated to fluorescein isothiocyanate (FITC) and employed as a platelet marker in whole blood samples as described previo~sly.~'*~"
Flow qtometric analysis We immediately placed 50 p L of whole blood (patient samples) or of PC in 1mL of 2-percent paraformaldehyde in phosphate-buffered saline and incubated it for 1 hour at 4°C. We than washed samples with Tyrode's buffer and incubated them in the standard fashion for 20 minutes at 22°C with a saturating concentration of biotinylated S12 and FITC-anti-GPIb, as described prcviou~ly.~~ The sample was washed again, and 20 pL of phycoerythrin (PE)conjugated avidin (Becton Dickinson Immunccytometry Systems) was added for 20 minutes' incubation. PEanjugated avidin specifically binds biotiny-
'IRANSFUSION
1991-Vd. 31. No. 5
411
ACTIVATION OF STORED PLATELETS
lated S12 to produce PE fluorescence on the surface of cells expressing GMP-140. We again washed the samples in "yrode's buffer and analyzed them in a flow cytometer (FACScan, Becton Dickinson Immunocytometry Systems). The instrument was Set to measure forward light scatter (i.e., a measurement of particle size), and FITC and PE fluorescence. We calibrated the instrument for forward light scatter by using 2- and 10-km beads (Duke Scientific Corp., Palo Alto, CA) and for fluorescence by using calibration beads (Calibrite, Becton Dickinson Immunocytometry Systems). We identified platelets in whole blood samples by setting the cytometer to detect FITC fluorescence; that is, only GPIbpositive particles were measured as platelets. We also performed gating by using the light scatter profile of platelets to separate them from the larger red and white cells, as detailed previo~sly,~'.~~ and from any platelet fragments. The level of background PE fluorescence was determined on a daily basis using PEconjugated avidin alone, without biotinylated S12. We periodically labeled ADP-stimulated whole blood samples and analyzed them in a manner identical to that of the study samples to confirm that the S12 antibody was functional. Analysis of 10,OOOplatelets was performed, and the percentage of activated platelets was determined by the proportion of platelets that expressed PE fluorescence higher than background (Fig. 1). Although an 80-percent decline in GPIb fluorescence was noted with maximal platelet activation, the threshold for FITC detection was low enough to include all platelets. Dual platelet labeling with GPIb and GPIIbma showtd no platelets which had entirely lost GPIb as a consequence of storage. control studits with GPIIbAIIa and S12 labtling shaved similar percentages of activation as did those with GPIb and
Four Day Old
S12 labeling. We noted no GPIb negative platelet-sized cells, which confirmed that only platelets were analyzed.
Results GMP-140 q m s i o n in blood bank-stored PCs The relationship of GMP-140 expression-that is, the percentage of platelets that are activated-and the length of platelet storage is shown in Fig. 2. There is an obvious increase in platelet activation with storage; when compared to freshly drawn whole blood, platelets in storage were significantly activated (p
